What public sector innovators have to say about using data to solve real-world problems
Summary of our roundtable discussion with seasoned innovators on the challenges and opportunities of using data and analytics to improve public services.
What public sector innovators have to say about using data to solve real-world problems
An increasing number of local government organisations are waking up to the potential of data and analytics to help improve their delivery of services and their decision making. Some are quite advanced on their journeys, with several projects already under their belt; others are just testing the water. Given the pressure on public services to meet spiralling demand in the face of tighter budgets, it’s in everyone’s interests that these early experiments are able to scale up to bring much wider impact to the sector.
So, what have we learned so far from the advanced practitioners? And what advice can we give those just starting out?
On 16 October 2017, Nesta and Mastodon C hosted a roundtable with 17 public sector innovators and data experts to explore this topic. This short summary captures the key ideas, challenges, and advice raised during that event. It is not a blow-by-blow account of what was said, but is instead organised under the key messages that emerged, presented below.
There is growing evidence of success
It was clear from the discussion that the idea of making smarter use of data has moved far beyond the hypothetical. Data is already being used to realise real savings and improve services. Participants offered examples of their work, which have led to measurable and significant impact.
For example, HMRC Connect links data from a range of government and corporate sources to create a profile of a taxpayer’s total income in order to spot irregularities. Since its launch in 2008, it has collected £3 billion in additional revenue. Meanwhile the London Borough of Camden’s Resident’s Index automates the matching of local citizen data, lessening the burden on frontline staff who spend less time searching for information and creating savings across services. In housing alone, the borough is on track to save £3.4 million by using the system to identify illegal sublets. Other London boroughs, including Brent and Barking and Dagenham have developed comparable solutions and are achieving similar benefits.
In short, though many in local government remain sceptical of data and its potential to drive change, early movers are already able to demonstrate meaningful results. Innovating with data is difficult work, but, as a number of our participants noted, the benefits can be well worth the effort.
It takes time to find the right problem to tackle with data
Not all problems lend themselves to a data-driven solution. A key area of the group’s discussion focused on how local authorities can determine which of their challenges are suitable for this approach. There was consensus that simply bringing people together without a specific agenda - even a highly motivated cross-section of staff and leadership - is unlikely to surface the kind of use-cases where data can offer impactful and actionable insight.
Many organisations might also be tempted to start by trying to marshal all the data they can on a topic, to see what insights it could provide. Most of our roundtable participants advised against this approach, as it is not likely to generate interesting or impactful outcomes, or save more time and effort compared to starting out with a clearly defined problem.
Participants suggested that it is first worth considering the kind of solutions that data analysis can provide, which include:
- Finding anomalies
- Prioritising cases based on risk or need
- Identifying early signs of risk
- Surfacing new insights through visualisation
- Streamlining operations and decision making
- Optimising resource deployment
From here, our discussion focused on the importance of starting with a problem that is both clearly understood by the organisation (i.e. they know why it happens or persists) and is meaningful to leadership.
Many organisations know the broad areas they wish to improve. However, what often poses a challenge is drilling down to a level of specificity needed to enable actionable insights. To get here, “continual and structured access to end users” is needed, in the words of one participant. This is necessary to obatin a greater understanding where data can usefully support the decision-making process and optimise performance. To make this possible, it is incumbent on leadership to ensure enough of their staff’s time (including frontline workers) is freed up to provide necessary guidance on relevant data-driven projects.
For data to be impactful, organisational cultures need to change
Data projects are often designed in a way that doesn’t take into account organisational culture. In the words of one participant, “you can’t just do data to an organisation,” meaning that data-driven projects require strong engagement with people, their motivations, and processes to succeed. A number of our roundtable participants therefore emphasised that ‘data projects’ should avoid labelling themselves as such, and instead focus on the outcome they’re meant to achieve. The data, they advised, should be framed as just one of the levers used to support the desired outcome, and a means rather than an end.
Data-driven projects require strong engagement with people, their motivations, and processes to succeed.
Organisations must also be mindful of the fact that “data doesn’t start neutral in people’s minds”, as noted by another participant. Some may view data initatives with a degree of suspicion or as something that could take up a lot of time without necessarily generating an equal amount of value.
Participants suggested a number of strategies to counter these perceptions and build confidence in new methods. To begin, they advised using words like ‘insights’ and ‘useful knowledge’ in place of ‘data’ in communication with staff to avoid alienating people. Secondments or co-location between frontline staff (or other end users) and analytics teams were also deemed particularly important for fostering trust, learning and stronger analysis. For instance, one participant gave an example of working alongside a social worker who validated all early output generated by his analysis, which not only served to develop a more accurate tool, but also to create trust and support from end users. Mastodon C urged that care must be taken to empower end users in this way, so as to minimise concerns or scepticism concerning the role of data in their work.
Lastly, as with any project involving cultural change, leadership has a significant role to play. Many participants noted that executive and senior management must “create the necessary space for experimentation”, give “explicit permission to break things”, and grant staff the “freedom to fail”. These are critical components of innovation that can only come from leadership and cannot be bought in, our participants stressed.
A better approach to procurement is required
Data and technology projects in the public sector have a tendency to overpromise and underdeliver, as one roundtable participant reminded the group. There are many reasons for this, one of which has to do with how such projects - often complex and highly novel - are procured by the sector.
Based on their experiences, our participants offered a few suggestions for how to approach procuring analytics and other data products. First, organisations should avoid specifying all requirements up front, given the high degree of uncertainty and complexity inherent in these projects. Second, a few participants emphasised the importance of building in a proof-of-concept (POC) phase. POCs provide strong benefits for both the organisation and the supplier. They build confidence and support by starting small, and allow for greater experimentation by giving the supplier permission to take some early risks. At the same time critical problems, but also opportunities, are revealed earlier in the process, when there is more time to intervene and adjust course.
Proof-of-concepts build confidence and support by starting small, and allow for greater experimentation by giving the supplier permission to take some early risks.
Finally, one participant suggested that a ‘bake-off’ approach - in which two or more vendors are selected to develop POCs and the best design wins the contract - could also increase chances of project success. In all, there was agreement that procurement should be less prescriptive and allow for quicker experimentation and iteration, especially in the early phases of a project.
The right balance between data protection and data innovation is hard to find
Leadership discussions on data currently tend to lean heavily toward protection and privacy concerns, rather than opportunities for innovation. As one participant pointed out, for many in local government, data simply has a “bad brand”, influenced by negative media coverage of high profile security breaches. Incoming data legislation, the GDPR, is also stoking worries over meeting higher compliance and documentation standards for personal information held by the public sector.
But what if instead of generating more concern and risk aversion to data-driven projects, the GDPR were to have the opposite effect? As one participant suggested, the GDPR actually presents a significant incentive for local government and other public sector bodies to bring their siloed data together, since doing so would make compliance with new rules easier. If they have to spend time and money on compliance, they might as well use those resources to solve some of the big challenges they have faced with data at the same time.
Unfortunately, it was feared that few organisations are likely to embrace the GDPR in this way, especially those where the potential of data to positively affect people’s lives or save money remains poorly understood. Without supportive leadership to shift the narrative and bring focus to opportunities and successes, however, progress in this area will be hard won.
There are common technical challenges to be overcome
Local authorities wishing to make smarter use of data are likely to come up against a common set of technical challenges. Among these, our discussion highlighted poor data quality (stemming in large part from a lack of standardisation, such as inconsistent recording of address data) and IT silos as some of the biggest roadblocks.
These barriers in particular make it difficult to join up data across an organisation and perform impactful analysis. They may also persist because of a general lack of understanding on the data architecture (the rules and models that define how data is collected and used in an organisation) needed to support analysis, as one participant expressed.
It is important to note however that several councils have already successfully addressed these challenges, and there is much to be learned from their experiences. Equally, it was suggested that the NHS, which has developed data standards to facilitate analysis, could serve as inspiration. Finally, as one participant underscored, it is worthwhile just getting started with the data on hand, even if incomplete or messy, as this will surface important gaps that must be addressed.
Some participants also commented that local authorities could be doing more with the technology they already possess to extract insights from their data. For example, the London Borough of Barking and Dagenham is maximising the use of its GIS tool to optimise bulky waste collection and is already looking for other similar applications, like scheduling graffiti clean-up. Again, leadership is needed to recognise and support opportunities like this, as well as to prioritise data architecture initiatives that will serve as the building blocks of analysis.
Scaling what works remains a challenge
Despite growing evidence and examples of how local government is using data to save money and improve outcomes, learning from one another and scaling what works remains a challenge for the sector.
Frustratingly, as one roundtable participant pointed out, a high degree of scepticism persists on whether proven approaches will work locally. There is a strong ‘not invented here’ view across the sector. Equally, open source tools and technologies, which are used in the development of various analyses and apps, are not being exploited as widely as we’d expect. Innovative collaboration platforms such as CC2I, a co-funding platform that allows public organisations to co-develop and obtain new tools at a fraction of the cost of going it alone, could also be leveraged to a much greater degree.
...leadership that understands the opportunities on offer and is able to move the organisation towards data innovation, is essential.
In order to scale what works, our participants agreed the sector needs to create better visibility and engagement around proven and promising approaches. This could be partially achieved by smaller efforts, for example through the use of more creative communication channels, such as Slack. New incentives and funding models - for example an ‘app store’ for local government - are also needed to promote more experimentation and uptake of existing solutions. As with all other key issues raised throughout the discussion, leadership that understands the opportunities on offer and is able to move the organisation towards data innovation, is essential.
Nesta is invested in supporting local and public sector authorities to learn from one another and innovate with data. In addition to organising events, workshops and roundtables like this one, we will also soon publish a Data Practice Guide, which will provide actionable advice for making smarter use of data in public organisations. With the support of the Local Government Association, Nesta is also developing a Data Maturity Tool to help local authorities assess their strengths and weaknesses across different areas, including data management, skills, governance and infrastructure.
In the meanwhile, to share the latest ideas and connect with colleagues working on data innovation in the public sector, we encourage you to join Mastodon C’s recently launched UK Data Connect Slack channel. To join, please email [email protected], and you will receive the link.
Mastadon C and Nesta wish to thank our roundtable participants for sharing their experiences and insights:
Alex Dewsnap, Divisional Director for Strategic Commissioning, LB Harrow
Bruce Durling, CTO and Co-founder, Mastodon C
Chris Lowry, Policy and Improvement Officer, Sheffield City Council
Deven Ghelani, Director, Policy in Practice
Eddie Copeland, Director Government Innovation, Nesta
Emma Rourke, Director Public Policy Analysis, ONS
Fran Bennett, CEO and Co-founder, Mastodon C
Hilary Simpson, Founder, Sleuth Cooperative
Maryvonne Hassall, Assistant Director for Digital Strategy, Aylesbury Vale District Council
Nevena Dragicevic, Programme Manager, Nesta
Niraj Dattani, Councillor, LB Harrow
Paul Hodgson, GIS & Infrastructure Manager, Greater London Authority
Paul Copping, Chief Innovation Officer, Digital Greenwich
Pye Nyunt, Insight Hub Manager, LB Barking and Dagenham
Rachael Tiffen, Head of Counter Fraud, CIPFA
Ravi Gogna, Principal Consultant, BAE Systems
Sarah Dougan, Chief Analytical Officer, LB Islington
Shawn Jhanji, Business Development, Mastodon C
Simon Bullmore, Marketing, Mastodon C
Steve Skelton, Strategic Head, Stockport Council
Toby Eccles, Development Director, Social Finance
Vicky Sargent, Client Services Director, Boilerhouse